1. Immunology of CML stem cells: Non-cycling, quiescent CD34+ CML cells are more resistant to TkI. We addressed the question of whether this small population of true leukemia stem cells (also characterized by the CD34+CD38-Lin-CD90+ phenotype), which may represent the reservoir of relapsing leukemia, are susceptible to immune attack by T cells and NK cells. To assess the potential of leukemias associated peptide vaccines to eliminating CML, we measured WT1, PR3, NE and PRAME expression in CD34+ progenitor subpopulations in CML patients and compared them with mHAgs HA1 and SMCY. All CD34+ subpopulations expressed similar levels of mHAgs irrespective of disease phase, suggesting that in the SCT setting, mHAgs are the best target for GVL. Surface expression of WT1 protein in the most primitive hematopoietic stem cells in AdP-CML suggest that they could be targets for WT1 peptide-based vaccines, which in combination with PRAME, could additionally improve targeting differentiated progeny, and benefit patients responding suboptimally to TkI, or enhance GVL effects in SCT patients. We next performed functional studies examining the effect of NK cells on cycling and quiescent CD34+ CML cells. Purified NK cells obtained from normal donors or from CML patients were expanded in vitro. Bortezomib significantly enhanced autologous NK cells cytotoxicity on cycling and quiescent non-cycling CD34+ CML cells by 20-40%. The increased sensitivity to autologous NK-cytotoxicity correlated with increased expression of tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) receptors DR4 and DR5 on the surface of CD34+ quiescent cells, and was reversed by blocking TRAIL. These findings support further development of bortezomib as an adjunct treatment with adoptive transfer of allogeneic expanded NK cells in association with SCT. 2) NK cells in myeloid malignancies: We studied SCT patients and patients with AML to correlate functional markers on NK cells with outcome of treatment, and freedom from relapse in patients transplanted for CML. We studied the relationship between the donors KIR genotype and transplant outcome. In a multivariate analysis, studying established factors predictive of relapse after SCT, and exploring known associations of KIR-mismatching with recipient HLA class I molecules, we found that only two factors emerged as predictive for relapse: patient disease risk group, and donor KIR type. KIR 2DS1, 3DS1 and 2 DL5A were favorable for sustained remission. This KIR genotype effect was entirely restricted to patients with AML: good risk patients with favorable KIR had a 0% relapse rate, compared with a 32% relapse rate if the donor lacked favorable KIR genotype. In high relapse risk patients, relapse rate was 30% for donors with favorable versus 70% for donors with unfavorable KIR. Future laboratory studies will explore how NK KIR polymorphism affects functional NK-AML target interactions. 3) Immune profiling in AML: To further explore NK and T cell function in AML in collaboration with Dr Stephen Strickland, (Vanderbilt University Hospital, Nashville TN) we are studying the immune profile of patients with AML at presentation, after remission induction, and at relapse. We explore the hypothesis that at presentation immune function is compromised by the leukemia through one or more mechanisms that suppress antileukemia responses. We monitor immune recovery and the circulating cytokines that drive lymphocyte recovery, characterizing the NK and T cell phenotypes in remission, studying NK function (cytotoxicity against standard targets and autologous blasts) and T cell function (frequency and cytotoxic characteristics of T cells recognizing LSA peptide mixtures including WT1, neutrophil elastase, Aurora kinase, MAGE 3 and PRAME). These studies will be repeated on patients who relapse to identify whether relapse is due to leukemia escape from immune control or whether immune control of residual disease is defective. Findings will be correlated with patient outcome to identify favorable immune recovery patterns that protect against relapse. Ultimately these studies should identify protective immune mechanisms in AML that inform rational immunotherapy (vaccination, blocking suppressor cells, cytokine treatment, NK or T cell adoptive therapy). 4) MDS: Patients with MDS carrying the trisomy 8 (tri-8) chromosome are particularly susceptible to immunosuppression.The ability to distinguish MDS cells from normal hematopoietic progenitor cells by the FISH technique and the frequent coexistence of mixed populations of normal and tri-8 cells in the marrow permitted experiments to determine the specificity of T-cell mediated suppression. We used T-cell colony inhibition tests to show that tri-8 progenitors were more susceptible to T-cell cytotoxicity that normal progenitors. However while tri-8 cells express apoptotic markers they appear to be less susceptible to complete apoptosis because of overexpression of anti-apoptitic proteins survivin and cyclin D2. We showed that tri-8 MDS progenitors overexpress WT1 and are susceptible to immune attack from WT1 specific CTL. These findings underpin the observation that patients with tri-8 MDS can respond to IST but retain a hematopoietic tri-8 population.. We next plan to identify the peptide epitopes recognized by cells in tri- 8 with a view to finding new class I and II peptide epitopes naturally recognized by CD4 and CD8 T cells from MDS patients to broaden the range of known immunogenic peptides of WT1. Incorporating these new peptide antigens in peptide cocktails could improve the efficacy of WT1 vaccines. MDS and immunosuppressive treatment: Because of the toxicities associated with ATG and CSA, we investigated with the use of alemtuzumab as a single agent in MDS. Based on prognostic criteria previously defined patients judged likely to respond to immunosuppression were selected. Thirty-two patients received alemtuzumab 10 mg/d intravenously for 10 days. Seventeen (77%) of 22 evaluable MDS intermediate-1 and four (57%) of seven evaluable intermediate-2 patients responded to treatment with a median time to response of 3 months. Four of seven evaluable responders with cytogenetic abnormalities before treatment had normal cytogenetics by 1 year after treatment. Five (56%) of nine responding patients evaluable at 12 months had normal blood counts, and seven (78%) of nine patients were transfusion independent. These results showed that alemtuzumab is safe and active in MDS and may be an attractive alternative to ATG in selected patients. 5) LGL: Previous studies showed that some patients with LGL respond to CSA. To explore whether more intensive immunosuppression would be more effective at restoring hematopoiesis, patients with LGL were given alemtuzemab using the schedule used in the MDS trial. Fifteen of 19 patients responded rapidly. All patients had a sustained and prolonged lymphopenia but without developing opportunistic infection. There was a massive reduction of the LGL clones in responders when evaluated at 3 and 6 months. Relapse was associated with resurgence of the clone. Leukapheresis samples prior to and 6 months after treatment with alemtuzemab will be used to identify changes in the lymphocyte repertoire in responding and non-responding patients and to define the mechanism of myelosuppression by LGL cells. Selected LGL samples whose CD8 T cells are over 95% clonal will be used to screen a combinatorial peptide library to identify the cognate peptides and attempt to identify the target protein (collaboration with Dr M Wooldridge, Cardiff, UK). Understanding how the LGL T cell suppresses myelopoiesis and erythropoiesis may shed light on the mechanism of suppression of both normal and leukemic cells by cytotoxic T cell populations.